HIGH-FIDELITY MODELS FOR COAL

COMBUSTION:

TOWARD HIGH-TEMPERATURE OXY-COAL

FOR DIRECT POWER EXTRACTION

XINYU ZHAO

UNIVERSITY OF CONNECTICUT

DANIEL C. HAWORTH1, MICHAEL F. MODEST

2, JIAN CAI

3

1THE PENNSYLVANIA STATE UNIVERSITY

2UNIVERSITY OF CALIFORNIA, MERCED

3UNIVERSITY OF WYOMING

SPONSORED BY NATIONAL ENERGY TECHNOLOGY LABORATORY

1

HIGH-TEMPERATURE

OXY-COAL COMBUSTION

Environmental

cleanup

equipment

Ash H2O SO2

Air in

Air

separation

unit

Combustor

N2 out

O2 in

Coal in CO2

capture

CO2

flue

gasSteam

turbine

Work

steamMHD

toppings

DC

CO2

flue

gasSteam

generator

Steam

turbine

steam

2

~2800 K

Challenging factors:

• Highly-concentrated chemically-reactive and radiatively-

participative species

• Shift of heat transfer patterns

• Change of modeling priority (devolatilization versus surface-

reaction)

OBJECTIVES

Explore the flow, chemistry, and heat transfer, and their

interactions through high-fidelity models.

3

Physics Proposed models

Turbulent flow field RANS-based model (LES)

Gas-phase chemistry Detailed chemistry models

Turbulence-chemistry-

radiation interactions

Transported PDF models

Radiative heat transfer P1 model (Photon Monte

Carlo)

Spectral properties of

gas/particle

k-distribution considering

CO2, H2O, CO, and

particles.

Char surface reaction

model

CBK model with CO2, H2O,

O2

Devolatilization model CPD model/ Two-rates

model with fitted parameter

A SYSTEMATIC APPROACH HAS BEEN ADOPTED

IMPLEMENT AND VALIDATE PMC-LBL FOR HIGH-T OXY-CH4

fuel jetoxygen jet

axis

furnace outlet

furnace

walls

18 14

8

unit: mm

oxygen

annulus

fuel jet

wall

3440

N. Lallemant, F. Breussin, R. Weber, T. Ekman, J. Dugue, J. M. Samaniego, O. Charon, A. J. Van Den Hoogen, J. Van Der Bemt, W. Fujisaki, T.

Imanari, T. Nakamura and K. IINO. Flame Structure, heat transfer and pollutant emissions characteristics of oxy-natural gas flames in the 0.7-1 MW

thermal input range. Journal of Institute of Energy, 73, pp. 169-182

4

Flame A: S. M. Hwang, R. kurose, F. Akamatsu, H. Tsuji, H. Makino and M. Katsuki. Application of optical diagnostics techniques to a

laboratory-scale turbulent pulverized coal flame. Energy Fuels (2005), 19, 382-392

FlameB: M. Taniguchi, H. Okazaki, H. Kobayashi, S. Azuhata, H. Miyadera, H. Muto, T. Tsumura. Pyrolysis and ignition

characteristics of pulverized coal particles. ASME. Vol. 123. pp. 32-38. 5

Flame AFlame B

A SYSTEMATIC APPROACH HAS BEEN ADOPTED

IMPLEMENT AND VALIDATE TRANSPORTED PDF – COAL SOLVER

OUTLINE

Methods

Chemistry and turbulence-chemistry interactions

Radiative heat transfer

High-fidelity models in coal combustion

Conclusions

6

METHOD

7

• notional particle tracking

• mixing

• reaction

• radiation

notional

particles• continuity

• momentum

• energy

• pressure

• k, ɛmean flow

densitymesh cells

notional

particles

photon bundles

Transported composition PDF method Photon Monte Carlo method + LBL

𝐼𝑏𝜂 ≠ 𝐼𝑏𝜂( 𝑇)𝑆𝑐 ≠ 𝑆𝑐( 𝑇) Turbulence-chemistry-radiation interactions

METHOD

Coal parcels in one

cell (solid phase)

Enthalpy and

mass sources for

each individual

species (+/-)

PDF notional particles

in one cell (gas phase)

8

FinallyTransported PDF/PMC/coal solver

The coupling model is a challenging aspect.

TYPICAL SUB-MODELS

Turbulence: standard k-epsilon model with adjusted Cɛ1

Chemical mechanisms

GRI-Mech 2.11

Radiative heat transfer

P1 radiation with gray gas and particles

P1 with k-distribution (Cai et al.)

Mixing models

Euclidean minimum spanning tree model (EMST)

Variable CΦ

Chemical acceleration

In situ adaptive tabulation (ISAT) (parallel)

Devolatilization

two-rates model

single-rate model

modified single-rate model

Surface reaction

diffusion-kinetic-control model

oxy-char combustion model **

*Mehta, R. S., Haworth, D. C., Modest, M. F. An assessment of gas-phase reaction mechanisms and soot models for laminar atmospheric-

pressure ethylene-air flames. Proc. Combust. Inst. 32, 2009, 1327-1337.

** Murphy, J. J., and Shaddix, C. R. Combustion kinetics of coal chars in oxygen-enriched environments. Combust. Flame. 144 ,2006, 710-729

9

Initialize RANS code

Solve coal parcels position,

velocity, energy transfer

Solve gas-phase

momentum, equivalent

energy, pressure

equations

Steady

state

Move particle in

physical and

composition space

Compute

equivalent energy

source term

Spectral

PMC

Transported PDF (Fortran)FV RANS (OpenFOAM in C++)

END

YES

NO

Add gas-phase particles

evolved from coal

Initialize PDF particle

codes

SOLUTION ALGORITHM

10

OUTLINE

Methods

Chemistry and turbulence-chemistry interactions

Radiative heat transfer

High-fidelity models in coal combustion

Conclusions

11

TCI IN HIGH-TEMPERATURE OXY-CH4

COMBUSTION

12

mean temperature

mean CO mass fraction

The effect of turbulence-chemistry interactions is reflected

in the flame structure.

TCI IN TURBULENT COAL COMBUSTION

OH PLIFComputed YOH 13

OUTLINE

Methods

Chemistry and turbulence-chemistry interactions

Radiative heat transfer

Advantage and difficulties of using high-fidelity

models

Conclusions

14

RADIATION IS DOMINANT IN THE HIGH-T

OXY-CH4

FLAME

x= 1.42m

mean temperature mean XCO2 mean XCO

15

SPECTRAL PMC MODEL CAN PREDICT

ABSORPTION COEFFICIENTS IN THE HIGHLY

PARTICIPATIVE ENVIRONMENT.

C. Yin, L. A. Rosendahl, S. K. Kær. Chemistry and radiation in oxy-fuel combustion: a computational fluid dynamics

modeling study. Fuel. 90(7), pp. 2519-2529.

HiTemp2010

database

prediction:

~ 2.

16

TURBULENCE-CHEMISTRY-RADIATION

INTERACTIONS ARE INTENSE ONLY IN THE

FLAME CORE, NEAR THE NOZZLE.

TRI

term 1 term 2 term 3 17

THE LABORATORY-SCALE FLAME IS

OPTICALLY-THIN. RADIATION HAS HIGHER

INFLUENCE TO LARGER PARTICLES.

18Courtesy of Jian Cai

gas Solid 1, Solid 2, Solid 3,

OUTLINE

Methods

Chemistry and turbulence-chemistry interactions

Radiative heat transfer

High-fidelity models in coal combustion

Conclusions

19

THE ASSUMPTION OF EQUILIBRIUM CHEMISTRY

CAN BE EXAMINED BY THE FINITE-RATE MODEL.

𝑫𝒂 =𝝉𝒇

𝝉𝒄𝑫𝒂𝒗𝒐𝒍 =

𝝉𝒇

𝝉𝒗𝒐𝒍

fast devolatilization

fast chemistry

H2O

𝜏𝑐ℎ𝑒𝑚 > 𝜏𝑣𝑜𝑙

CO

20

𝑓𝑑𝑒𝑣𝑜𝑙 =𝑚𝑑𝑒𝑣𝑜𝑙

𝑚𝑝𝑟𝑖𝑚𝑎𝑟𝑦 + 𝑚𝑝𝑖𝑙𝑜𝑡 + 𝑚𝑑𝑒𝑣𝑜𝑙𝑓𝑠𝑢𝑟𝑓 =

𝑚𝑠𝑢𝑟𝑓

𝑚𝑝𝑟𝑖𝑚𝑎𝑟𝑦 + 𝑚𝑝𝑖𝑙𝑜𝑡 + 𝑚𝑑𝑒𝑣𝑜𝑙 + 𝑚𝑠𝑢𝑟𝑓

21

THE MIXTURE FRACTIONS OF

DEVOLATILIZATION AND SURFACE REACTION

CAN BE RECONSTRUCTED FROM THE RESULTS.

-0.99 -0.8

pt1 pt2

0.320.9

pt3 pt4 22

THE ASSUMPTIONS USED IN THE MIXTURE

FRACTION BASED METHOD COULD BE TESTED

USING HIGH-FIDELITY MODELS.

DIFFICULTIES IN USING HIGH-FIDELITY MODELS

Coal parcels in one

cell (solid phase)

Enthalpy and

mass sources for

each individual

species (+/-)

PDF notional particles

in one cell (gas phase)

Model 1: distribute cell-level mass and energy source weighted by

notional particle mass (mp)

Model 2: distribute cell-level mass and energy source weighted by

notional particle temperature (Tp)

Model 3: distribute cell-level mass and energy source weighted by

reactivity exp(-C/Tp) (C is a constant) 23

24

MORE VALIDATION WITH EXPERIMENTS OR

HIGHER-ORDER MODEL IS NECESSARY.

CONCLUSIONS

• A transported PDF model for coal combustion using finite-rate chemistry has been built. Components of the model has be validated through a hierarchy of experimental configurations.

• The spectral photon Monte-Carlo method implemented in this work can capture the nongray effect of the high-temperature oxy-combustion environment naturally.

• Turbulence-chemistry-radiation interactions can also be captured by the model without additional effort. The interactions are extremely important for pollutants prediction (CO, NO, and soot).

• The high-fidelity models developed in this work can be used to guide the development of simpler models.

• LES model will be coupled with PDF method to properly predict the particle location.

25

POSSIBLE FUTURE WORK

26

CO2

O2

27